Return-Path: Received: from eastrmfepo201.cox.net ([68.230.241.216]:46565 "EHLO eastrmfepo201.cox.net" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1752272Ab1F2ScP (ORCPT ); Wed, 29 Jun 2011 14:32:15 -0400 Date: Wed, 29 Jun 2011 13:32:35 -0500 From: Tom Haynes To: peter.staubach@emc.com Cc: tdh@excfb.com, SteveD@redhat.com, linux-nfs@vger.kernel.org Subject: Re: [PATCH 1/1] Remove warnings on various platforms Message-ID: <20110629183235.GA6422@adept.internal.excfb.com> References: <1309299723-31539-2-git-send-email-tdh@excfb.com> <5E6794FC7B8FCA41A704019BE3C70E8B8285F299@MX31A.corp.emc.com> <20110629172858.GA6258@adept.internal.excfb.com> <5E6794FC7B8FCA41A704019BE3C70E8B8285F3CB@MX31A.corp.emc.com> Content-Type: text/plain; charset=us-ascii In-Reply-To: <5E6794FC7B8FCA41A704019BE3C70E8B8285F3CB@MX31A.corp.emc.com> Sender: linux-nfs-owner@vger.kernel.org List-ID: MIME-Version: 1.0 On Wed, Jun 29, 2011 at 02:16:24PM -0400, peter.staubach@emc.com wrote: > Hmmm. Yes, make sense. > > That said, I think that it would be better to attempt to do it the right way and then let people tell us when it isn't quite right yet. We should eventually be able to come to a set of changes which are right, for everyone. Okay, I'll do that and resubmit. > > Compiling on Linux, Solaris, and FreeBSD should be a good start. Perhaps Olga or someone else could help with the Windows compiles. I suspect that we could find someone with access to AIX or HP/UX or ... :-) > > Perhaps we could also consider cleaning up tests.init? The lists of commented out options seems quite unwieldly to me. Do we really need all of that stuff, especially for the much older systems? Yeah, I've been wondering if modern tools like autoconfig would work here. > > Thanx... > > ps > > > -----Original Message----- > From: Tom Haynes [mailto:tdh@excfb.com] > Sent: Wednesday, June 29, 2011 1:29 PM > To: Staubach, Peter > Cc: tdh@excfb.com; SteveD@redhat.com; linux-nfs@vger.kernel.org > Subject: Re: [PATCH 1/1] Remove warnings on various platforms > > On Wed, Jun 29, 2011 at 07:53:19AM -0400, peter.staubach@emc.com wrote: > > Hi. > > > > For some of these new casts, wouldn't be it better to simply declare the variable to be the correct type to start with instead of casting some other type? For example, the socklen_t variables? > > > > It seems funny that all of those XDR routines need to be casted as well. > > > > Thanx... > > > > ps > > I tried to go with the most innocent changes I could. My main concern > is lack of access to different platforms to do sanity compiles > and unit testing. > > -- > Tom Haynes > ex-cfb > -- Tom Haynes ex-cfb